Glioblastoma multiforme (GBM), the most common primary brain tumor in adults, carries a dismal prognosis. Clinical trials with non-replicating retroviral vectors for gene therapy of GBM showed inadequate gene transfer, without therapeutic benefit. Hence we developed a unique strategy with retroviral replicating vectors (RRV) for more efficient & tumor-selective delivery of prodrug activator (`suicide') genes, without spread to normal tissues. In intracranial glioma models, this enabled tumor-localized activation of an anti-cancer drug from a systemically administered non-toxic prodrug, achieving long-term survival without systemic myelotoxicity. Supported by prior NINDS U01 funding, an optimized RRV expressing the yeast cytosine deaminase suicide gene (RRV-CD, `Toca 511') was developed, and first-in-human Phase I trials for recurrent high-grade glioma have shown radiographic complete responses and significantly increased survival compared to contemporaneous external controls, leading to FDA `Breakthrough Therapy' designation, and a pivotal Phase III trial that is now underway. Ideally, this approach may be augmented by using RRV to deliver 2 or more suicide genes in combination, analogous to combination chemotherapy, but generated within the tumor itself, without systemic myelotoxicity. As viral packaging capacity limits us to a single suicide gene per virus, we developed a second RRV delivering another suicide gene, Herpes simplex thymidine kinase (TK), which activates anti-herpetic prodrugs such as ganciclovir by phosphorylation, and acts as a super-antigen. Also, as RRV encoated with the same envelope compete for binding to the same cell surface receptors, we encoated RRV-TK with Gibbon ape leukemia virus (GALV) envelope to avoid such interference with the current clinical vector RRV-CD, and enable combined use of both viruses. However, since GALV-encoated RRV may show unwanted transduction of hematopoietic cells, here we propose studies to validate the safety and efficacy of this new RRV. As the GALV envelope binds a cellular receptor conserved on human and rat (but not mouse) cells, for intracranial tumor models to evaluate tumor transduction, biodistribution, and genotoxicity, we propose to establish both (a) human glioma xenografts in immunodeficient mice with humanized hematopoiesis, and (b) syngeneic gliomas in immunocompetent rats. To mitigate potential genotoxicity, we also propose to evaluate a novel strategy to restrict RRV by incorporation of the target sequence for microRNA miR-142-3p, which is known to be highly abundant in hematopoietic cells. In the R21 Phase, we will first examine replicative efficiency and selectivity of the newly developed GALV- encoated RRV in glioma vs. hematopoietic cells in vitro (Aim 1), and optimize the above in vivo models (Aim 2). In the R33 Phase, we will employ these models to evaluate tumor selectivity and transduction efficiency of GALV-encoated RRV (Aim 1), and to evaluate its efficacy for suicide gene therapy and activation of anti-tumor immunity (Aim 2). If successful, we will establish a lead candidate for further clinical translation of GALV-encoated RRV-TK, as a single agent by itself and/or combined with RRV-CD, through the CREATE Program.